If you're playing the long game, because B2B buyers make decisions based on familiarity and trust, this is good news. But you'd be forgiven for missing that, given the volume of hot takes that followed. New hacks. New fears. And, as ever, a rush to share the latest ‘quick wins’ while declaring the old playbook obsolete!
LinkedIn published a full technical breakdown of the update on their Engineering blog. It’s worth reading before anyone else's commentary shapes your thinking.
Here’s the update in their own words:
“While the Feed has long been AI-powered, recent LLM advances gave us the opportunity to rethink what's possible. That's why we're rolling out a new advanced ranking system, powered by LLMs and GPUs, that better understands what a post is actually about and how it relates to a member's evolving interests and career goals."
The algorithm doesn't define success. Your content does.
So what does LinkedIn actually value? The algorithm filters for relevance, evaluating what your post is about, who it's likely to matter to, and whether you're the kind of voice that shows up consistently on that topic. What the new LLM-powered system changes is the precision of that matching, not the underlying logic.
LinkedIn's algorithm update is designed to be more adaptive to evolving user interests, rather than being guided purely by historical engagement data. In practical terms, the old system rewarded familiarity. If someone had engaged with your posts before, they were more likely to see them again.
In my training, I always emphasised engagement as the key to building future visibility. That principle still holds, but this new system now broadens the opportunity, matching content to current interests means your posts can reach relevant audiences beyond those who already know you.
There's also a clear signal on what's being deprioritised. Engagement bait, posts that prompt users to "Comment 'Yes' if you agree”, or posts that feature a video with no relation to the accompanying text, is being actively filtered out. Recycled thought leadership posts that add little in terms of substance or insight will also be downranked. A welcome shift (and long overdue).
If your activity has been built around gaming engagement rather than earning it, this update is a genuine problem. For everyone else, it's an improvement.
The update is good news for anyone producing quality content: when industry news breaks and relevant posts gain traction, the updated system surfaces them within minutes, not hours or days.
But here's what most people are missing
The feed algorithm is one piece of a much larger picture of change. According to Semrush, LinkedIn is now the second most cited domain across AI search tools, sitting ahead of Wikipedia, major news publishers, and other social platforms (Reddit is currently occupying first place).
Semrush analysed 325,000 prompts across three AI tools, ChatGPT Search, Google AI Mode, and Perplexity, identifying 89,000 LinkedIn URLs cited in responses. On average, 11% of AI responses reference LinkedIn content.
Why does this matter? AI responses tend to mirror the meaning of original LinkedIn content closely, your framing, your positioning, your expertise can surface directly inside an AI-generated answer that a potential buyer is reading right now. That's a different kind of visibility than feed reach, and it's one most B2B marketers haven't yet accounted for.
So what earns that kind of citation? Educational, original content published consistently, long-form articles and substantive posts that share practical knowledge rather than promotional messaging.
The research also found that roughly 75% of cited authors post frequently, and about half have over 2,000 followers, so follower count matters less than you might expect. Authoritative content from smaller accounts still breaks through.
One nuance worth flagging for those managing both a brand presence and a personal profile: not all AI tools draw from the same sources. Perplexity tends to cite Company Pages most often, while ChatGPT Search and Google AI Mode more frequently surface content from individual creators. That's a compelling data point to have in your back pocket when making the internal case for investing in both, it's not either/or, it's both/and.
What this means for you
The latest algorithm update and the AI visibility data are pointing in the same direction. Stop optimising for the feed. Start optimising for building and maintaining trust.
That means posting consistently on the topics where you have genuine expertise. It means writing with clarity of thought, not volume of words. It means publishing original insight rather than recycled takes. And it means treating LinkedIn less like a broadcast channel and more like a body of work. one that both human readers and AI platforms are actively drawing from.
A word on consistency. Playbooks might tell you to post three times a week, always on Tuesday mornings, and between 8 and 10am. That guidance has its place, but it misses the point. Consistency that matters is consistency of topic and perspective, not frequency for its own sake. Showing up once or twice a week with genuine insight on the things you actually know will always outperform daily posting that says nothing in particular. The algorithm, and your audience, will notice the difference.
Start by auditing your last 90 days of LinkedIn activity against a simple test: does each post reflect genuine expertise, answer a question your audience is actually asking, and come from a consistent point of view? If the answer is mostly yes, this update works in your favour. If not, that's where your energy is better spent.
And don't overlook your profile. The system uses everything LinkedIn knows about you, your headline, summary, skills, and experience, to understand who you are and what you should be associated with. If your profile still reflects where you were three years ago rather than where you are now, the algorithm is working with outdated information. A profile that clearly signals your expertise and focus area makes everything else work harder.
